Innogrid Launches Next-Generation Virtualization Platform 'Openstackit 3.5' to Bolster AI Infrastructure Support
TECHWORLD ·
✦ AI Summary
Innogrid announced on the 3rd the launch of its next-generation private cloud solution, 'Openstackit 3.5.'
'Openstackit 3.5' was refactored based on the latest OpenStack version, 'Gazpacho,' and strengthened high availability, resource optimization, migration and monitoring functions, along with large-scale virtualization infrastructure operating features.
It also expanded GPU management functions to support efficient resource operation in AI training and inference environments, and is pursuing a strategy to expand beyond replacing existing VMware environments into infrastructure that can support cloud-native and AI workloads.
Innogrid announced on the 3rd the launch of its next-generation private cloud solution, 'Openstackit 3.5.' 'Openstackit 3.5' aims to improve stability and operational efficiency in large-scale virtualization environments, while also targeting support for AI workloads.
'Openstackit 3.5' has been completely refactored on the latest OpenStack version, 'Gazpacho.' Reflecting operating experience from existing customer environments, the company strengthened high availability functions, resource optimization functions, migration functions and monitoring functions, while also adding features needed to operate large-scale virtualization infrastructure.
It also applied a 'maintenance mode' that moves VMs on nodes scheduled for maintenance to other nodes in advance and restores the VMs after the work is completed. In addition, the company added an 'instance optimal placement' function, which supports VM distribution based on analysis of resource usage by compute node and helps ease resource concentration on specific nodes.
The company also strengthened fault-response functions. When a specific node fails, it provides 'instance HA,' which automatically transfers VMs running on that node to healthy nodes, helping ensure service continuity. It also supports 'hot plug,' which allows CPU and memory to be expanded for VMs in operation without interruption, backing continuous operations.
It also includes features to improve migration and management efficiency in large-scale virtualization environments. By applying multithread-based parallel live migration, it enables multiple data transfer tasks to be handled simultaneously, and it has also upgraded monitoring and metering functions for checking the status and usage of physical and virtual resources.
To address demand for AI infrastructure, the company expanded GPU management functions. It supports GPU resource provisioning, GPU resource allocation, real-time monitoring of GPU resources and GPU resource scheduling, helping improve the efficient operation of computing resources in AI training and inference environments.
Innogrid is pursuing a strategy to expand Openstackit 3.5 beyond a replacement solution for existing VMware virtualization environments and into infrastructure that can support cloud-native and AI workloads.
To that end, Innogrid operates a 'VMware modernization program,' whose purpose is to support existing customers' phased transition. The support method involves pre-analysis of customer infrastructure and workloads, followed by migration, DR, infrastructure transition and support at each operating stage.
Going forward, the company plans to link Openstackit with its own technologies, including CMP, PaaS and DevOps. The goal is to support everything from the transition of existing virtualization infrastructure to the buildout of cloud-native environments and the operation of AI services within a single technology framework.
This direction is tied to Innogrid's 'From xPU to AI Platform' strategy. The strategy envisions connecting a wide range of computing resources, including GPU, neural processing units (NPU) and central processing units (CPU), to GPU PaaS, AI CMP and AI PaaS, with the aim of expanding its technology scope into an AI platform managed within a single control environment.
Ban Il-kwon, head of the Cloud Infrastructure Center at Innogrid's R&D division, said that recent changes in the virtualization market show the competitive focus shifting away from replacing specific hypervisors and toward competition over infrastructure that can support stable operations, transitions and, in the future, GPU and AI workloads. He added that Innogrid aims to continue advancing virtualization and cloud infrastructure technologies to provide a technological foundation that enables stable transitions from existing environments and flexible, scalable expansion into AI infrastructure.
Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406489
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.